Default Output

The Options Object

The options object is optional and provides a number of controls to allow you to tailor the output to your needs. However, for general use it is recommended that all options are left to their defaults.

$3

String - valid options: 'freq' (default), 'binary', or 'percent'

N.B - You probably don't want to change this, ever.

Controls how the lexical value is calculated.

__Binary__ is simply the addition of lexical weights, i.e. word1 + word2 + word3.

__Frequency__ encoding takes the overall wordcount and word frequency into account, i.e. (word frequency / word count) * weight. Note that the encoding option accepts either 'freq' or 'frequency' to enable this option.

Another way to think of binary and frequency encoding is that 'binary' essentially sets all weights to '1', whereas frequency will generate a group norm. This is useful for predictive lexica, for example, when predicting age we want to use frequency encoding because we care about the actual number generated - i.e. the lexical value is the predicted age. Whereas, when predicting gender (see predictGender) 'binary' encoding is used because we don't care about the value, only whether it is above or below 0.

__Percent__ returns the percentage of total (non-unique) tokens matched against the lexicon in each category as a decimal, i.e. 0.48 = 48%.

$3

$3

$3

Number - accepts floats

Exclude words that have weights above the max threshold or below the min threshold.

By default these are set to positive and negative infinity respectively, ensuring that no words from the lexicon are excluded.

$3

Boolean - valid options: true or false (default)

The lexica contain intercept values, set noInt to true to ignore these values.

Unless you have a specific need to ignore the intercepts, it is recommended you leave this set to false.

$3

String - valid options: 'lex' (default), 'matches', or 'full'

'lex' (default) returns the lexical value, which is the predicted age.

'matches' returns an array of matched words along with the number of times each word appears, its weight, and its final lexical value (i.e. (appearances / word count) * weight). See the output section below for an example.

'full' returns an object containing the predicted age and the matches array. The keys are "age" and "matches".

$3

Number - valid options between 0 and 20 inclusive.

Number of decimal places to limit outputted values to.

The default is "undefined" which will simply return the value unchanged.

$3

String - valid options: 'lex' (default)', 'weight', or 'freq'

If 'output' = 'matches', this option can be used to control how the outputted array is sorted.

'lex' (default) sorts by final lexical value, i.e. (word frequency * word count) / word weight.

'weight' sorts the array by the matched words initial weight.

'freq' sorts by word frequency, i.e. the most used words appear first.

By default the array is sorted by final lexical value, this is so you can see which words had the greatest impact on the prediction - i.e. the words which lower the age the most appear at the beginning of the array, the words which increase the age the most appear at the end.

{'output': 'matches'} output example

Each match between the lexicon and your input is pushed into an array which contains: the term, the number of times that term appears in the text (its frequency), its weight in the lexicon, and its lexical value (i.e. (word freq / total word count) * weight) when 'freq' encoding is used).

By default the matches output array is sorted ascending by lexical value. This can be controlled using the "sortBy" option.